continuous-learning

To extract reusable patterns from Claude Code sessions automatically, configure the continuous-learning skill in your settings.json to evaluate completed sessions and save valuable interaction segments into a dedicated directory.

Session pattern detection analyzes completed Claude Code interactions to identify valuable debugging techniques, error resolutions, and project-specific conventions, building a growing library of learned skills for future use.

Yes, you can configure extraction sensitivity, session length thresholds, and pattern types to ignore. This customizable configuration allows tuning the pattern detection logic to match your specific software development workflow needs.

You can save error resolutions as skills by running automated pattern detection on completed session transcripts. The skill analyzes the interactions and extracts the resolution steps into a dedicated directory for quick future access.

Yes, session transcripts are required as input data. The skill analyzes these transcripts using configurable pattern detection logic to identify and save valuable interaction segments as reusable skills.

The best way to capture project-specific conventions is to automate pattern extraction using configurable detection logic on your code generation sessions, which continuously identifies and saves valuable techniques into a dedicated knowledge base directory.